Upload
dohanh
View
233
Download
3
Embed Size (px)
Citation preview
1
ICS 180: Digital Signal Processing
Aditi Majumder
Instructor
• Aditi Majumder
• Room: CS 352C
• Email: [email protected]
• Website: http://www.ics.uci.edu/~majumder/DSP/dsp.htm
2
The Course
• Fun course
• You will implement things and see results
• Digital Image Processing
• Do not worry about grades, enjoy the course
• Tell your friends
Course Load
• Learn about both 1D and 2D
• Programming Assignments mostly on 2D– Image Processing
• Written Assignments
• Midterm: 12 May (Wednesday)
• Finals
3
Logistics
• No late homework submission• No fixed office hours
– Send me email
• Grading Policy– Programming Assignments: 30%– Written Assignments: 10%– Midterm: 30%– Finals: 30%
Syllabus
• Signals and Linear Systems (0.5)• Convolution and Properties (2)• Fourier Transform and Properties (3)• Continuous Signal Processing (0.5)• Applications to Audio and Image
Processing (4)• Text: The Scientist and Engineer’s
Guide to Digital Signal Processing, Steven W. Smith
4
Hands On Experience
• Filtering (Low, high, and band pass)
• Edge Crispening
• Edge Detection
• Feature Detection
• Noise Cleaning
• FFT and Convolution
• Image Resampling
• Image Morphology
What is a signal?
• Any function dependent on a single or multiple variables
t
Am
pli
tud
e
1D: A = f ( t )
x
y
2D: I = f ( x, y )
x
y
z
3D: M = f ( x, y, z )
5
Origin of Signals
• Real life applications– Medical Images
– Seismic Vibrations
– Audio and Video
– Radar and Sonar
Signal Processing
• Science of processing signals– Data analysis
– Data compression
– Data Storage
– Data Retrieval
– Data Communication
– Special Effects
6
Applications
• Space– Photo Enhancement, Segmentation
• Medical– Diagnostic image analysis
• Media– Audio and video compression, Quality Control
• Military– Radar and sonar data analysis
• Scientific– Emergency monitoring and analysis
Analog Signals
• Both independent and dependent variables can assume a continuous range of values
• Exists in nature
7
Digital Signals
• Both independent and dependent variables are discretized
• Representation in computers
• Sampling– Discrete independent variable
– Sample and hold (S/H)
• Quantization– Discrete dependent variable
– Analog to Digital Converter (ADC)
Digital Signal
8
Sampled Signal
Digitized Signal
•Depends on number of bits•12 bits = 4095 levels• 0.0 ≤ Voltage ≤4.096•2.56 and 2.5601 TO 2560•Each level (LSB) = 0.001•Error ≤ ±½ LSB•Called Quantization Error
9
Digitized Signal
•Depends on number of bits•12 bits = 4095 levels• 0.0 ≤ Voltage ≤4.096•2.56 and 2.5601 TO 2560•Each level (LSB) = 0.001•Error ≤ ±½ LSB•Called Quantization Error
Quantization Error
• Usually like random noise
• Noise is present in most signal acquisition systems
• Random uncorrelated samples added to the original signal
10
Proper Sampling
• If the original signal can be reconstructed unambiguously from the sampled signal
Cycles/ Sample =Number of cycles per second
Number of samples per second
=Analog Frequency
Sampling Rate
Is it Proper Sampling?
• DC signal
• Freq = 0.0 x Sampling Rate
• Proper
11
Is it Proper Sampling?
• Freq = 0.09 x Sampling Rate
• Each sample covers 0.09 cycles
• Proper
Is it Proper Sampling?
• Freq = 0.31 x Sampling Rate
• Larger fraction of cycles per sample
• Proper
12
Is it Proper Sampling?
• Freq = 0.95 x Sampling Rate
• Much larger parts of cycles per sample
• Not Proper
• Aliasing
• Changes frequency and phase
Sampling Theorem
• Proper Sampling: At least one sample per half cycle
• Freq ≤ 0.5 x Sampling Rate
• Sampling Rate ≥ 2 x Frequency
• Nyquist Rate
13
Time (Spatial) Domain vs. Frequency Domain
• Any one dimensional analog signal can be represented as a linear combination of sine waves of different frequencies
1D Signal
• Example: Once scan line of an image
• Amount of each sine wave defined by its amplitude and phase